| || || || || || |
460 KARA-DONG AND THE SEARCH FOR HSÜAN-TSANG'S PI-MO [Chap. XIII
nor could our guides give any clear idea where we should find the ancient sites previously described as in its vicinity. Amidst the gradually rising dunes we soon left behind the last traces of old cultivation, dead trunks of Jigdas and Tereks, while the tamarisk-covered patches assumed more and more the appearance of isolated cones owing to incipient erosion of the unprotected ground around them.
As we plodded on, the villagers I had taken along for eventual excavations became more communicative. They professed never to have seen the sites we were bound for, but they were well aware of the legend concerning them. These same villagers had before shown a clear perception of the direct cause which had led to the abandonment of their old lands. All the greater was my surprise to find that the legend they now proceeded to tell me, of the ' old town ' marked by the remains in the desert beyond, was in all substantial points the same as that which more than twelve hundred years ago Hsüan-tsang had heard at 131-mo of the sand-buried city of Ho-lao-lo-chia. A holy man was treated with contempt by the inhabitants and refused water—as one version of the story asserted, because he had reproved them for unnatural crimes. He thereupon cursed the town, and foretold its approaching destruction. While they still mocked at his prophecy, sand began to rain from the skies, and continued for seven days and nights until the whole of the buildings were buried. Only seven pious people who had shown respect for the ulûgh izdum (` the holy man ') managed to save their lives through a curious device, which varies from Hsüan-tsang's story. The seven wise men are supposed to have clung to ropes fixed to a high pole after the fashion of a merry-go-round, still a popular diversion throughout Chinese Turkestan. Being whirled round and round by the raging storm they rose higher and higher above the ground while the sand accumulated, and thus escaped.
Similar stories, no doubt, are current throughout Eastern Turkestan of ruins buried in the Taklamakan, but it is of particular interest to note how the continuity of local tradition had here transferred to the remains of 31i-mo itself the legend which Hsüan-tsang heard at Fl-mo of a still earlier site. For I could safely recognize these remains in the extensive débris-covered area, portions of which we managed to trace in the course of the next two days. On that first evening our luckless guides dragged us aimlessly far out into the desert, until at last the weariness of animals and men and the difficulty of getting the caravan in the darkness over the rising dunes forced us to pitch camp without having come across any indication of an old site. During the night one of the guides deserted. The other, however, a timid young fellow whom Turdi, my desert factotum, kept under his eye, and encouraged by advice drawn from his own lifelong ` treasure-seeking ' experience, recovered his bearings, and setting out before daybreak succeeded in discovering one portion of the ` kône-shahr about three miles to the south-west.
Uzun-Tati—` the distant Tati ' 2, as local tradition appropriately designates this site—proved to consist of several extensive patches of ground, thickly covered with pottery fragments and similar small débris. The northernmost of those which I saw showed an area about one mile from north to south, and over a quarter of a mile across, surrounded by low dunes and tamarisk-covered sand-cones. A few of the latter rose also within the otherwise level débris-strewn area. The effects of long-continued erosion were visible everywhere. Only at relatively few spots little banks of loess showed foundations of mud-walls, nowhere more than a foot or two high. In some of these stumps of round wooden posts remained, with bundles of Kumush placed vertically between them, indicating a mode of construction closely approaching the modern one as observed in the dwellings of ` old Domoko'.
' Pronounced with the usual epenthesis, ta'ti, the second part of the name sounded here almost like téti.